AR Board Game Feel Study

ISEF Category: Technology Enhances the Arts

Subcategory: Games  ·  Difficulty: Advanced  ·  Setup: University Lab  ·  Time: Full Year

The Hook

A game can feel different even when the rules stay the same. A phone screen, a printed tile, and an AI opponent can change how players plan, react, and enjoy a match. That makes this project part game design, part computer vision, and part human behavior study.

What Is It?

This project asks a simple question with a lot of layers: how does an augmented reality board game feel compared with a digital-only version? In the AR version, you point a phone camera at printed AprilTag tiles, which are square markers that help a camera identify objects fast and accurately. A CNN, or convolutional neural network, reads the board state, and a Monte-Carlo Tree Search AI picks its moves.

Think of it like a chess set that can see itself. The board lives on your table, but the phone turns it into a smart system. You are not only testing whether the computer can recognize the pieces. You are also testing whether the mixed physical and digital format changes the player's sense of presence, fairness, clarity, and fun.

Why This Is a Good Topic

This is a strong science fair topic because you can measure more than one thing. You can test vision accuracy, AI decision quality, and player perception with surveys or play tests. That gives you real data, not just a demo. The project connects to game design, computer vision, and human-computer interaction, and you can learn how interface choices change the way people experience a game.

Research Questions

• How does an AR board game compare with a digital-only version in player-rated fun and immersion?
• What is the effect of board tile layout density on AprilTag detection accuracy?
• To what extent does camera angle change CNN-based tile segmentation errors?
• Which AI difficulty setting leads to the highest player engagement without lowering perceived fairness?
• How does a physical board plus phone interface affect move planning time compared with a screen-only interface?
• What is the effect of lighting conditions on recognition stability during play?
• To what extent does opponent move predictability change the player's rating of game feel?

Basic Materials

• Smartphone with a rear camera and tripod mount, if available.
• Printed AprilTag tiles or markers on matte paper.
• Flat game board base.
• Computer with internet access.
• Survey form tool for player feedback.
• Spreadsheet software for logging test results.
• Timer or stopwatch.
• Notebook for observation notes.

Advanced Materials

• Smartphone with manual camera controls.
• Laptop or desktop with GPU access, if available.
• Camera calibration target.
• Programmable lighting setup.
• Printed board variants with different marker densities.
• Python environment with OpenCV and a CNN framework.
• Game logging system that records moves, recognition confidence, and AI search outputs.
• Survey platform for structured player testing.

Software & Tools

• Python: Processes board images, runs analysis, and manages experiment logs.
• OpenCV: Detects markers, checks segmentation quality, and measures image stability.
• ImageJ: Measures visual contrast and marker visibility across board designs.
• Google Forms: Collects player ratings of fun, clarity, and fairness.
• JASP: Runs basic statistical tests and visualizes survey results.

Experiment Steps

1. Define the one experience variable you want to test first, such as board layout, lighting, or AI strength.
2. Separate technical performance from player experience so you can measure both recognition accuracy and game feel.
3. Build a baseline version, then plan a comparison condition that changes only one major design element.
4. Create a logging plan that captures board detection errors, move latency, and AI choices in the same session.
5. Design a player survey that asks about clarity, fairness, immersion, and fun using consistent rating scales.
6. Plan your analysis before you collect data, so you know which metrics will answer your question.

Common Pitfalls

• Testing board recognition under changing indoor light, which makes AprilTag detection look worse or better for the wrong reason.
• Changing the camera distance between trials, which alters marker size and breaks fair comparisons.
• Mixing up AI strength with AI speed, which makes players judge the game feel based on waiting time instead of move quality.
• Asking vague survey questions like "Was it fun?", which gives you data you cannot compare across players.
• Comparing AR and digital versions with different rules, which hides whether the interface itself changed the player experience.

What Makes This Competitive

A stronger version of this project ties human experience data to technical performance data. You can compare not just which interface players like, but why they like it, using recognition accuracy, move latency, and survey results together. You can also test a deeper design question, such as whether certain AI behaviors make a physical board feel more alive than a screen-only game. Clear controls, clean metrics, and a thoughtful analysis plan will matter more than flashy graphics.

Project Variations

• Test whether different tile densities change recognition accuracy and player comfort on the same board game.
• Compare AprilTag detection with another marker system, then measure whether one leads to fewer missed moves and better game feel.
• Study how human players rate fairness when the AI uses different search depths or response delays.

Learn More

• OpenCV Documentation: Learn marker detection and image processing basics in the official OpenCV docs.
• MIT OpenCourseWare Computer Vision: Find free lecture materials on image analysis and machine vision.
• PubMed: Search for review articles on human-computer interaction, player experience, and game immersion.
• IEEE Xplore: Search for papers on augmented reality games, computer vision, and player behavior.
• NIH PubMed Central: Read full-text open-access studies on interface design and user perception.
• Google Scholar: Search for recent papers on Monte-Carlo Tree Search, game AI, and mixed-reality game design.